(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Effect of parent-child interaction therapy on the verbalizations of behavior disordered children"

EFFECT OF PARENT-CHILD INTERACTION THERAPY 

ON THE VERBALIZATIONS OF 

BEHAVIOR DISORDERED CHILDREN 



By 

LAURA LYNN MEE 



A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL 

OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT 

OF THE REQUIREMENTS FOR THE DEGREE OF 

DOCTOR OF PHILOSOPHY 



UNIVERSITY OF FLORIDA 
1991 



ACKNOWLEDGMENTS 

The author wishes to express her appreciation to her 
committee members for their contributions and support of 
this work. The author wishes to particularly thank her 
chair, Dr. Sheila Eyberg, for the many long hours she has 
put into this study. Dr. Eyberg "s enthusiasm and 
encouragment sustained this author throughout this project, 
Finally, the author wishes to thank her family and husband 
for their support and patience. 






11 



TABLE OF CONTENTS 

page 

ACKNOWLEDGEMENTS ii 

LIST OF TABLES iii 

ABSTRACT v 

INTRODUCTION x 

Language Acquisition 3 

Social Development 17 

Behavior Disordered Children 19 

Parent-Child Interaction Therapy 24 

Specific Hypotheses 28 

METHODS 30 

Sub j ects ...!!! 30 

Observational Measure '.'.'.'. 31 

Procedure 35 

Child Speech Categories '.'.'.'. 37 

RESULTS 41 

Reliability !!!!!!!! 41 

Child Talk Variables Hypotheses ......!!!!!!! 46 

Total Speech Hypothese 53 

DISCUSSION 56 

Reliability !!!!!!!!! 57 

Child Talk Summary Variable Hypotheses ....!! 60 
Comparison of Total Amount of Speech of 

Mother-Child Dyads 64 

Implications for Treatment 6 6 

Future Directions , 68 

REFERENCES 69 

BIOGRAPHICAL SKETCH 74 



111 



LIST OF TABLES 



page 

Table One 

Category Priority Order for Coding Child Speech 36 

Table Two 

Reliability of Child Talk Categories 43 

Table Three 

Reliability of Summary Variables 44 

Table Four 

Interrater Agreement 45 

Table Five 

Pearson's Correlations by Summary Variables 47 

Table Six 

Negative Child Talk Means and Standard Deviations 49 

Table Seven 

Child Commands means and Standard Deviations 50 

Table Eight 

Child Questions Means and Standard Deviations 51 

Table Nine 

Child Playtalk Means and Standard Deviations 53 

Table Ten 

Child Praise Means and Standard Deviations 54 

Table Eleven 

Total Speech of Mother to Child 58 



iv 



Abstract of Dissertation Presented to the Graduate School of 

the University of Florida in Partial Fulfillment of the 

Requirements for the Degree of Doctor of Philosophy 



EFFECT OF PARENT-CHILD INTERACTION THERAPY ON THE 
VERBALIZATIONS OF BEHAVIOR DISORDERED CHILDREN 

By 

Laura Lynn Mee 

August, 1991 

Chair: Sheila M. Eyberg, Ph.D. 

Major Department: Clinical and Health Psychology 

Although there has been a substantial amount of 
research on the effect of the child's early environment on 
later deviant behaviors, there has been less research on the 
effect of the child's early language environment on 
subsequent deviant verbalization. Mothers who display 
deficient/ inappropriate verbal behavior in parent-child 
interactions have been found to provide poor verbal role 
models and may thus facilitate their child's development of 
such deviant verbal behaviors. 

Parent-Child Interaction Therapy (PCIT) is a program 
for the treatment of behavior problem children in which 
parents are taught positive communication skills as well as 
behavior management. Utilizing the Dyadic Parent-Child 



Coding System (DPICS) , it has been documented that maternal 
negative verbalizations decrease over the course of therapy. 
What was of interest in the present study, and had not been 
investigated before, were the changes in child 
verbalizations over the course of therapy. In the present 
study, DPICS was expanded to include child talk categories 
which were coded on 20 behavior problem preschool -age boys 
with their mothers in all three DPICS situations prior to 
and following PCIT. 

The basic hypotheses were (a) that over the course of 
the therapy program the child's negative verbalizations 
would decrease and pro-social verbalizations would increase 
and (b) that these language changes would be reflected in 
better give-and-take between the mother-child dyad and would 
lead to more similar amounts of verbalizations following 
treatment. The results suggest that the therapy program 
does indeed affect the child's verbalizations in a positive 
way. Results indicate significantly less negative child 
talk from these children following treatment. However, some 
speech categories changed in ways not predicted and the 
total amount of speech between the mother-child dyad did not 
become more similar following treatment as had been 
hypothesized. Further research in this area is needed to 
replicate these findings and expand them in other samples. 



VI 



INTRODUCTION 

Through-out the past century many different and at 
times opposing theories of language acquisition have been 
proposed. Theories ranging across many fields and many 
theoretical orientations have been expounded to explain 
this all important developmental area in children. In 
the early part of this century, theories focused 
exclusively on innate (biological) causes or on strict 
environmental causes of language acguisition, yet it is 
now accepted that no one theory is entirely correct. 
Today, many theories emphasize a contribution from the 
social environment of the child, especially focusing on 
the influence of the primary caretaker as a socializing 
agent. Snow (1979) asserts that research has demonstrated 
that the crucial elements of language acguisition are 
readily available through social interactions. 
Socialization and the language learning environment are 
now seen by many as key ingredients to the child 1 s 
language acguisition, development of a verbal repertoire, 
and subsequent language patterns. 

More recently, research has extended beyond language 
acquisition to encompass the child's development of 



2 

language skills (i.e. the ability to use language for 
adaptive purposes) through the early childhood years. The 
acquisition of language skills is crucial for the child's 
development of interactions with others that serve to 
enhance the child's knowledge base. One of the early 
researchers who emphasized the importance of the child's 
social interactions in many areas but particularly in the 
development of language skills was the Russian researcher 
Lev Vygotsky (1962, 1978). His research focused on the 
interactions between language and behavior in young 
children and how adults influence their development. 
Vygotsky theorized that it was directly through the 
social-interactive process with the parent that children 
acquired the ability to regulate their own behavior 
through verbal mediation. 

Although there has been a substantial amount of 
research on the effects of the child's early environment 
on later problem behaviors, there has not been this same 
interest in the area of early language models on 
subsequent negative child verbalization. It is well known 
that the child's environment can have long-lasting 
positive or negative effects on the child's actions or 
behavior (e.g. Patterson, 1982) . Of interest in the 
present study is the effect of the verbal environment on 
the child's verbal behavior. 



Parent-Child Interaction Therapy (PCIT) is a therapy- 
program formulated to treat psychological disturbances in 
pre-school age children (Eyberg, 1979) . In addition to 
changes made in overt behavior of the parent as well as 
the child, changes in maternal verbalizations are taught 
and have been observed (Eyberg & Robinson, 1982) . 
Specifically, decreases in maternal negative 
verbalizations and increases in positive verbalizations 
over the course of therapy have been attained. What has 
not been investigated, and is the topic of this study, is 
the subsequent effect on the child's verbalizations this 
therapy program generates. 

Language Acquisition 
As one of the most important aspects of child 
development in the first five years, language acquisition 
and subsequent continued development have been extensively 
studied. Many theories of language acquisition fall more 
or less along a continuum banded on one end by theories 
that hold language acquisition to be an innate 
developmental process and on the other by theories that 
strictly emphasize the actual verbal behavior as the focus 
for the study. Theories relating to these two extremes 
will be briefly discussed as will two theories that fall 
in a somewhat more middle ground and posit the influence 
of socialization and interaction with others in the 



4 
child's language learning environment. These last two 
theories, Piaget's and Vygotsky's, will be elaborated on 
as two examples of what Lewis and Cherry (1977) have 
termed the Interactionist model; theories that integrate 
the areas of cognition, language and socialization. There 
is increasing support for the central role social 
dimensions play not only in the acquisition but in the 
development of language in young children. 

Innate/Biological Theory 

The primary assumption of the biological or innate 
developmental theorists is that language acquisition most 
often follows a predictable maturational process 
(Lenneberg, 1967) , and knowledge of syntactic and 
phonological rules is innately available (Chomsky, 1965) . 

Four aspects of language acquisition as outlined by 
Lenneberg (1967) point to it being a maturationally 
controlled behavior. These four areas, that can be 
applied to many of the maturationally dependent processes 
of child development, will be elaborated in the specific 
context of language acquisition. They are the following: 

1) The Regularity of Onset 

Language acquisition has a "regularity in the 
sequence of appearance of given milestones, all correlated 
with age and other concomitant developmental facts" 
(Lenneberg, 1967, p. 126). The onset of speech occurs 



5 
through a series of important milestones that are reached 
in a fairly fixed sequence at predictable chronological 
ages (Lenneberg, 1967) . 

2) Relation of Environment to Age of Acquisition 
There is "evidence that the opportunity for 

environmental stimulation remains relatively constant 
throughout development, but that the infant/ child makes 
different use of such opportunities as he/she matures" 
(Lenneberg, 1967, p. 126) . 

3) Role of Utility in Language Acquisition 

There is an "emergence of the behavior either in part 
or entirely, before it is of any immediate use to the 
individual" (Lenneberg, 1967, p. 126). According to 
Lenneberg, rudimentary beginnings of speech such as cooing 
and babbling appear to be part of the innate repertoire of 
the human species. Research from the biological or innate 
theorists has often stressed the continuity between 
pre-verbal behavior and the subsequent development of 
language . 

4) Importance of Practice for Language Acquisition 
There is "evidence that the clumsy beginnings of the 
behavior are not signs of goal-directed practice" 
(Lenneberg, 1967, p. 126). It is Lenneberg's opinion that 
cooing and babbling do not represent "practice stages" for 
subsequent language. In accordance with the idea that 
language acquisition is one of many maturational processes 



6 
in which human proficiency appears to be innate, practice 
is not seen as something that is needed. 

Lenneberg's (1967) theory is just one representative 
of innate theories on language acquisition. It does 
however, convey the theoretical position of most innate 
theories, that being an emphasis on the maturational 
course and innate knowledge base of language acquisition. 
At the other end of the language acquisition continuum is 
the operant conditioning theory postulated by B. F. 
Skinner (1957). 

Environmental Theory 

The premise of Skinner's approach to language 
acquisition is that the basic terms and processes involved 
in the science of behavior are general and applicable to 
all behavior, including verbal behavior. The vocal 
operant, like any other operant in the individual's 
repertoire, will increase in frequency and will be 
sustained when the operant is consistently followed by 
some sort of reinforcement (Skinner, 1957) . Skinner 
(1957) viewed operant reinforcement as a way of 
predictably organizing verbal behavior. During the 
process of language acquisition, the more stringent 
criteria upon which reinforcement is usually contingent is 
substantially moderated. Any vocal response approximating 
the given societal criteria is reinforced. As the child's 



7 
vocal responses increase in quantity and quality, closer 
approximations are required. It is in this way that the 
standard and yet complex verbalizations of adulthood are 
attained (Skinner, 1957) . 

As in all types of operant conditioning, Skinner 
stressed that the focus of studies on language acquisition 
should remain on the actual verbal behavior, as opposed to 
how to get the child to display (emit) the behavior in the 
first place. Skinner (1957) stated "it is difficult, if 
not impossible, to discover stimuli which evoke specific 
vocal responses in the young child. In order to reinforce 
a given response we simply wait until it occurs" (p. 31) . 
According to Skinner, the behavior is what comprises what 
is to be explained rather than ideas, desires, intentions, 
physiological or chemical mechanisms (Winokur, 1976) . 
Skinner is critical of those theorists who state the cause 
of developmental change is an assumed internal maturation 
following the species-specific genetic clock (Thomas, 
1985) . 

Interactionist Theories 

Finally, theories for language acquisition that 
postulate an interaction between the self and others 
besides the individual will be briefly discussed. This 
section encompasses theories that look at language 
acquisition as one part of the overall development of the 



8 
child. Specifically, these theories emphasize 
socialization and cognition as primary concomitant aspects 
of development along with language acquisition. 

Lewis and Cherry (1977) describe a model of language 
acquisition that is unidimensionally related to cognition 
and socialization. Referred to as the Interactionist 
Model, this model assumes that language development 
interacts with social and cognitive development; however, 
all three remain as discrete areas. The language 
development theories of both Piaget and Vygotsky fall 
under this Interactionist Model (Lewis & Cherry, 1977) . 

Vygotsky' s theory (1962) stated that although 
language and cognition develop separately, they influence 
each other and interrelate in various ways. Piaget 's 
developmental theory of cognition, language and 
socialization was based on the notion of a dependence on 
the development of underlying sensorimotor intelligence or 
a quantitative increase in information-processing capacity 
(Pascual-Leone, 1970) . Thus, the separate realms of 
language, social and cognitive development in Piaget 's 
theory are interrelated to a certain extent in that all 
are dependent upon an underlying development of primary 
processing abilities. Both Vygotsky' s and Piaget 's 
theories on language acquisition will be further 
elaborated at this point. 



9 
Piaget's definition of language was more restrictive 
than both Lenneberg's (1967) which extended down to 
beginning with the infant's first coos and babbles, and 
Skinner's (1957) global definition of language as another 
operant behavior. Piaget defined language as only those 
vocalizations occurring after the capacity for mental 
representation was present (beginning at approximately age 
two) . Piaget described mental representation as the 
ability of the child to produce mental images or symbols 
that represent objects and relationships. 

With development, the child is increasingly less 
dependent on direct physical behavior to know something. 
Instead, to learn about things, the child relies more and 
more on the mental representation of that action (Thomas, 
1985) . For Piaget, language was acquired then, as part of 
what he labeled the pre-operational period (age 2-7 
years). The most important aspect of Piaget's 
pre-operational language acquisition was his view that it 
had two distinct functions or phases. The first he 
labeled the egocentric and the second the socialized 
function (Brainerd, 1978) . 

Piaget's egocentric speech was speech that was either 
not directed to another or was not easily understood by 
others (Berk, 1985) . The term egocentric conveyed 
Piaget's belief that the young child was incapable of 
differentiating his or her own perspective on events from 



10 
the perspective of others and therefore operated in an 
egocentric fashion (Brainerd, 1978) . By comparison, 
Piaget's socialized speech was seen as communicative. He 
regarded speech that attempted to persuade listeners or to 
provide them with some useful item of information as clear 
examples of socialized language (Brainerd, 1978) . Piaget 
believed that the shift from egocentric to more socialized 
speech in the child was a product of social interaction 
with both peers and adults. 

Although both Vygotsky and Piaget worked in the area 
of language development at about the same time, Vygotsky' s 
(1962) work was not translated from the original Russian 
until the 1960s whereas Piaget had been translated into 
English in the 1930s. Vygotsky' s work has since gained 
some of the same attention in the Western world that has 
been accorded Piaget's work for over 40 years. 

Vygotsky' s work in language did not go into detail 
regarding early acquisition. Vygotsky believed that the 
child's parents played an important role in this learning 
and he chose to focus on child language in children ages 5 
to 11 years. Vygotsky 's research on children's language 
development between these ages emphasized its usefulness 
to the child and the important role adults play. 

As it compares to Piaget's theory for children 
between those ages, Vygotsky 's position is in direct 
contrast to Piaget's notion that egocentric speech serves 



11 

no useful function for the child. Vygotsky, instead, saw 
what he termed private speech as the child's vocalizations 
of self-guidance because the young child cannot yet think 
entirely in his head (internalized thought) as older 
children and adults are able to do. Vygotsky 's position 
suggested that private speech denoted the child ■ s 
dependence on verbal stimuli to promote internal thought 
processes and to mediate or regulate behavior (Vygotsky 
1962, 1978). Where Piaget viewed egocentric speech as 
useless, Vygotsky and Luria (1930) researched and reported 
on the importance of private speech for subsequent task 
success, thereby showing its usefulness to the language 
learning child. 

Vygotsky (1978) hypothesized that a child's 
self -regulatory private speech originated in the 
directions or verbal guidance given to the child by an 
adult. Self-regulation, then, was seen to begin as a 
social-interactive process between the child and the 
adult. Vygotsky believed that the child would then 
gradually come to regulate his or her own behavior by 
employing these internalized, learned verbal dialogue 
strategies which were originally initiated by the adult 
(Pellegrini, 1980) . Private speech was represented as the 
mid-point between the points where regulation is suggested 
by the adult and the final form where regulation is 



12 
covertly exercised in the form of inner speech (thought) 
by the child. 

The recent increased interest in Vygotsky's work has 
directed attention back to one of his theoretical 
cornerstones: the influence of adult's speech on the 
child's private speech. A central point of Vygotsky's 
theory lies with his assumption that a child's private 
speech is formed/molded through parental interchange, 
meaning that the child's self-regulating private speech 
actually originates from the directions given the child in 
a social-interactive process with the parent (Pellegrini, 
1980) . From this framework evolved the notion of the zone 
of proximal development. 

The zone of proximal development has been described 
as the distance between the child as the recipient of an 
instructional social interaction and the level at which 
the child functions independently (Wertsch, 1979) . Within 
this zone, the use of the tools and techniques of the 
society are introduced and practiced in a social 
interaction with more experienced members. This teaching 
is most typically done through verbal exchanges, and 
therefore, language skills profit greatly from these 
social interactions. 

Whereas Piaget postulated that growth for the child 
was a solitary venture in which others could not help, 
Rogoff and Wertsch took their cue from Vygotsky and 



13 
surmised that "growth was a collective responsibility and 
language one of the major tools of that collectivity" 
(Rogoff & Wertsch, 1983, p. 96). It has been said that 
the acquisition and development of language expands the 
child's communicable knowledge, by enabling him/her to go 
beyond personal experience and learn from the society at 
large (Oleron & Lorion, 1977) . With the continual 
development of an ever increasing vocabulary beyond the 
acquisition stage, the young child achieves an ability to 
learn about and explore the environment. 

This teaching is an essential part of the child's 
development of social understanding and social 
interaction. The child's increase in socialization can 
then be used by the parent to assist the child in 
attending to specific characteristics or patterns of 
behavior that will then elicit interpretations and 
expectations about the social world (Durkin, 1986) . The 
transmission of these societal norms from parent to child 
involves an interactional process that relies heavily on 
language. Children, as developing social beings, "use 
language as one means of acquiring, organizing and 
consolidating social knowledge and as the primary medium 
of social activity" (Durkin, 1986, p. 205). The teaching 
of societal knowledge from one generation to the next is 
well represented by the typical parent-child dyad. 



14 
The person most intimately involved in this 
instructional social interaction (guidance) of the young 
child is usually the child's mother. The child must 
initially depend on the mother for guidance and direction 
before being able to understand social relationships, 
react appropriately to a given situation or problem solve 
in an independent manner. The developmental progression 
from other-regulation to self-regulation is an important 
component of the mother-child dyad. Rondal (1985) found 
that mothers uniquely adapt their speech to the 
developmental level of their child and that their speech 
changes as the growing child's linguistic ability changes 
over time. That is, the mother adjusts her speech 
(guidance) as the child's zone of proximal development in 
various areas advances. 

Some of the first descriptions of mothers' speech to 
young children were undertaken in the late 60s to dispute 
the then popular view of language acquisition as largely 
innate (e.g., Chomsky, 1965; Lenneberg, 1967) and 
occurring independently of the child's environment. Snow 
(1979) has shown through numerous studies that information 
about the "structure of communicative systems" is 
available to children through social interactions. From 
this research on maternal speech, three points were 
articulated by Snow (1979) . These points were as follows: 
(a) language acquisition is the result of the process of 



15 
interaction between mother and child which begins early in 
infancy, with the child making as important a contribution 
as the mother (these interactions are crucially important 
to the cognitive and emotional development of the child as 
well as to language development) ; (b) language 
acquisition is guided by and is the result of cognitive 
development; and (c) the study of the mother's simplified 
speech to the young child is as important a topic for 
language acquisition research as those of syntax or 
phonology. 

These key arguments refuting innate theorists were 
echoed by Bruner (1975) when he stated that language 
development "entails learning how to produce and receive 
utterances in situational contexts. .. .We see this 
developing out of the pre-linguistic communication 
established in the first year of life" (Bruner, 1975, p. 
17) . Bruner (1975) further stated that as the child 
grows, he/she learns that other people's signals have 
communicative value through interactions (e.g. playing 
games) with mother. Additionally, Snow described maternal 
utterances to their young children as simplified to 
children up to three years of age with few changes from 
subject-verb forms. Snow used these data to hypothesize 
that children learn word order rules through modeling of 
these initially simplified phrases during social 
interactions (Snow, 1972; Snow & Ferguson, 1977). 



16 
Researchers such as Bruner, Vygotsky, and Snow tend to 
refute the Lenneberg and Chomskian view that language 
acquisition is innate and instead are part of a number of 
language researchers who accept the concept of a factor of 
parental (social) influence on the child's language 
acquisition and subsequent continuing development 
(Vygotsky, 1978; Snow & Ferguson, 1977; Snow, 1972). 

Rondal (1985) argued that a parent's understanding of 
language development in general, and of their own child in 
particular, including verbal productions and receptive 
reactions, helps to determine the specific level of 
adaptation in the parent's speech to the child. Others 
(Peterson & Sherrod, 1982) have even gone so far as to 
suggest that language delay in children who are otherwise 
intellectually average may be due to the detrimental 
linguistic style of their mothers. 

Research by Peterson and Sherrod (1982) with language 
delayed children has shown that the amount of semantically 
unrelated conversation by the mother in these dyads was 
greater than for normals. The increase in unrelated 
conversation indicates a serious lack of rapport 
(synchrony) between the child and his or her mother and 
may be why the language delayed children are having 
difficulty in that what they hear is often unrelated to 
the ongoing action. The mothers in the Peterson and 
Sherrod (1982) study made fewer approving comments and 
were less responsive to the child's speech. Because the 



17 
children in that study had no inherent organic or 
intellectual deficits, it was speculated that the delay in 
language acquisition might have been the result of the 
poor linguistic interactions with the mother. It has been 
found that mothers who display deficient/ inappropriate 
verbal behavior in dyadic interactions provide poor role 
models which may thus facilitate their child's development 
of similar deficient verbal repertoires (Patterson, 1986) . 

Social Development 
For many years the focus of socialization studies was 
on the parent's impact upon the child. The child was seen 
as a tabula rasa to be molded by parental influence — a 
passive being (Lytton, 1980) . More recently the impact of 
society on the dyad as well as the child's influence has 
been looked at more closely. Emphasis has shifted to the 
reciprocal interactions between parent and child. 
Socialization is now viewed as a complex process of 
interaction between the child and others in the child's 
social network. This interaction is thought to have 
bidirectional effects, changing both the child and those 
with whom the child interacts (Damon, 1983) . That is, the 
child's own influence helps determine the social 
experiences and thereby changes the socialization process 
from one generation to the next. Because of this 
bidirectionality, socialization is no longer viewed as 
just the transmission of cultural values and rules (Mead, 



18 
1963) but the integration process of the child into a 
social world that, in fact, has been influenced and 
changed by the presence of the child. 

Damon (1983) illuminated two major functions of 
social development as essential for the individual's 
adaptation to a society. These two are the integration 
into the society's culture, and the individuation into the 
individual's personal identity. At times these two forces 
appear to oppose one another and yet there are strong 
connections between the two. As early as 1902, Baldwin 
wrote that children come to know themselves only as a 
conseguence of social interaction with others. And Durkin 
added that as social beings most of our skills are 
"acquired and consolidated in the course of interactions 
with others" (1986, p. 204). Through verbal interactions 
with adults and more capable peers, children develop and 
continually refine their understanding of their social 
environment . 

Socialization, then, is viewed as a complex process 
of interaction between the child and others. Most notably 
during the child's early development, this socialization 
involves interactions between the child and parent in 
which language allows the child to gain entrance into the 
social framework. Durkin stated that the child's 
"understanding and operating within the social world is 
closely tied to understanding the language of others, to 
articulating one's own social reasoning, needs and goals, 



19 
and to participation in joint discourse" (Durkin, 1986, 
p. 232). It is through these social interactions with 
parents that the child develops an initial sense of social 
competence (Damon, 1983) . And it is likewise through 
these interactions that misconceptions and asocial 
tendencies can develop. 

Lack of social competence can have deleterious 
effects on the child's subsequent integration into the 
family and the society at large. The deficits in 
socialization skills can result in significant 
difficulties. Indeed, socialization problems are 
presently hypothesized to play a major role in the 
initiation of conduct problem behaviors in children 
(Patterson, 1982, 1986). As adults, not only are these 
children more likely to continue to engage in antisocial 
behavior, but are also at an increased risk for 
psychiatric impairment, poor occupational adjustment, low 
educational attainment, marital distress/disruption, and 
poor physical health (Kazdin, 1979) . 

Behavior Disordered Children 
As described in DSM-III-R (APA, 1987) , the essential 
feature of Oppositional Defiant Disorder consists of a 
"pattern of disobedient, negativistic, and provocative 
opposition to authority figures" (APA, 1987, p. 57). 
Current diagnostic criteria include demonstrating the 
behavior problems for at least six months and displaying 



20 
at least five of the following: loses temper, argues with 
adults, actively defies or refuses adult reguests or 
rules, deliberately does things that annoy others, blames 
others for own mistakes, is touchy or easily annoyed by 
others, is angry or resentful, is spiteful or vindictive, 
swears or uses obscene language. 

The prevailing theoretical formulation concerning the 
development and maintenance of conduct-problem behaviors 
has emphasized the primacy of familial socialization 
processes (Patterson, 1982, 1986). Patterson emphasizes 
the coercive, or controlling, nature of this type of 
behavior and has developed a theory to account for its 
development and maintenance. Negative reinforcement plays 
a particularly important role in Patterson's formulation, 
in that coercive behavior on the part of one family member 
(parent or child) is reinforced when it results in the 
removal of an aversive event being applied to another 
family member. As this 'training' continues over long 
periods, significant increases in the rate and intensity 
of these coercive behaviors occur as children are 
reinforced for engaging in aggressive and oppositional 
behaviors. Furthermore, the child also observes his or 
her parents engaging in coercive responses, and this 
provides the opportunity for modeling to occur (Patterson, 
1982) . 



21 
Socialization Deficits in Behavior Disordered Children 

Patterson's research with behavior problem children 
has led him to the conclusion that "in effect, these 
children's socialization appeared to be severely slowed" 
(Patterson, Reid, Jones & Conger, 1975, p. 4). Data from a 
study by Patterson in the homes of boys labeled as 
aggressive indicated that the subject's behaviors were 
more typical of much younger children. Patterson went on 
to state that these children who were labeled as socially 
aggressive often had not learned how to obtain and/or 
maintain positive peer relationships, were more often 
rejected by peers, freguently received more punishment 
than nonaggressive children, tended to acquire academic 
skills more slowly and were unlikely to simply outgrow 
their aggressive behavior over time (Patterson et al., 
1975) . 

The idea that behavior disorders in elementary school 
years and social competency/development may be linked, in 
part, stems from the adult literature suggesting that 
adult psychopathology and social competency are related 
(Bellack & Hersen, 1979; Kazdin, 1985). The question then 
is what factors might account for continued behavior 
disorders over time? Bellack and Hersen (1979) concluded 
that social competency, or the lack thereof, may influence 
a child's long-term prognosis. Children diagnosed with 
these behavior problems are likely to exhibit similar 
patterns of behavior into adulthood if left untreated, and 



22 
have an increased likelihood of engaging in delinquent or 
criminal behavior (McMahon & Forehand, 1988) . 

Similar to the inappropriate and deviant actions that 
are modeled for the conduct problem child by the parents, 
the verbal behavior models that the parent displays for 
the child are also deficient (Patterson et al., 1975). 
Subsequently, conduct problem children may acquire from 
their early language learning experiences in interaction 
with their mothers the same inappropriate language 
patterns that the parents demonstrate. Although there has 
been a substantial amount of research on the effects of 
the child's early environment on the development of 
conduct problem behaviors, there has been less research on 
the effects of the child's early language environment on 
subsequent negative and/or socially delayed 
verbalizations. It is likely, according to Vygotsky's 
theory (1962, 1978), that these children who display 
inappropriate language patterns did not receive adequate 
verbal teaching to internalize problem solving and 
socially appropriate verbalizations. A study exploring 
differences in the verbalizations of preschoolers with 
conduct problem behavior and a matched comparison group 
will next be discussed. 



Comparison of Verbal Behaviors in Conduct Problem and 
Comparison Children 

In a study by Forster (1984) , 20 conduct problem 



23 
children and their mothers were observed in the 
child-directed interaction (CDI) situation of the Dyadic 
Parent-Child Coding System (DPICS) (Eyberg & Robinson, 
1983) . These dyads were compared to a group of 20 normal 
mother-child dyads. The DPICS was expanded to code all 
child verbalizations exhaustively, as it does for parent 
verbalizations, by applying the parent verbalization 
category codes to the child's speech. Results indicated 
no differences between conduct problem children and 
comparison children in a combined category labeled 
"deviant behavior," which included both vocal (cry, yell, 
sass, whine) and nonvocal (destructive) negative 
behaviors. However, conduct problem children used 
significantly fewer positive verbalizations (praise, 
questions) , more playtalk (speaking through toy 
characters) , and among boys, more commands with their 
mothers than the comparison children. 

As noted by Forster (1984), the finding of no 
differences in deviant behaviors during CDI may be due to 
the generally low rates of negative behavior emitted by 
all children in situations where the mother does not 
exercise control, and is consistent with previously 
reported lower rates of deviant behaviors in conduct 
problem children in CDI than in the Parent-Directed 
Interaction (PDI) situation of DPICS (Eyberg & Robinson, 
1982). Further support for the possibility that negative 
verbalizations may be situation specific is provided by 



24 
data on parent verbalizations showing that the ratio of 
criticism to praise is higher during the PDI than the CDI 
situation for both behavior problem and normal mothers 
(Aragona & Eyberg, 1981) . 

Forster's (1984) finding that conduct problem 
children used fewer positive verbalization (praise, 
questions) during CDI than comparison children suggests 
the possibility that these children have socialization 
deficits in positive verbalization that are more pervasive 
but less noticed because they are not disruptive to 
others. The category of Child Commands has not previously 
been considered as a deviant behavior, but the finding 
that conduct problem boys use commands more often in CDI 
suggests that it may function as a negative verbalization 
in the form of "bossy" behavior. The finding that 
playtalk was higher in conduct problem than comparison 
children was an unpredicted finding in the Forster (1984) 
study, and it is unclear if this was a spurious finding or 
represents a form of indirect communication in conduct 
problem dyads. A treatment program that focuses on 
changing verbal as well as nonverbal behaviors will next 
be discussed. 

Parent-Child Interaction Therapy 
Parent-Child Interaction Therapy (PCIT) was 
formulated to treat psychological disturbances in 
preschool age children (Eyberg, 1979) . Based in large 



25 
part on the two-stage model of parent-training developed 
by Constance Hanf (1969) , PCIT is unigue in its emphasis 
on more traditional play therapy technigues and 
problem-solving skills training (Eyberg & Boggs, 1989) . 
Specific aspects of the original Hanf therapy have been 
maintained including directly coaching the parent while 
he/she interacts with the child, and explicit training in 
differential social attention and a time-out procedure 
(Eyberg & Boggs, 1989) . 

There is a wide and steadily increasing body of 
literature indicating that the parent-child interaction 
model of behavioral parent training is a successful means 
of reducing conduct problem behaviors in children. It has 
been shown to reduce noncompliance and deviant behaviors 
in the clinic (Eyberg & Robinson, 1982) and to generalize 
from the clinic to the home (Peed, Roberts, & Forehand, 
1977; Baum & Forehand, 1981). Maintenance of behavioral 
improvements in treated children and of behavioral skills 
in parents has been demonstrated at one year following 
completion of parent-child treatment (Baum & Forehand, 
1981) . PCIT has further been documented to decrease 
maternal negative verbalization and increase positive 
verbalizations over the course of therapy (Eyberg & 
Robinson 1982) . 



26 
Description of Therapy Package 

Parent-Child Interaction Therapy (PCIT) is a 
treatment program in which parents are taught positive 
communication skills as well as behavior management 
(Eyberg & Robinson, 1982) . Parent-Child Interaction 
Therapy involves direct interaction between the parent and 
the child which comprises the majority of each therapy 
session. Directions and feedback are provided immediately 
to the parent regarding the ongoing interaction. A 
bug-in-the-ear microphone is used to "coach" parents to 
respond with positive verbalizations to prosocial child 
behaviors or interactions and to use ignoring or time-out 
techniques following all negative child behaviors or 
interactions (Eyberg & Boggs, 1989) . 

Parent-Child Interaction Therapy consists of two 
phases: Child-Directed Interaction (CDI) and 
Parent-Directed Interaction (PDI) . Each phase has several 
components designed to change both the child's and the 
parent's behavior (Eyberg & Boggs, 1989). The primary 
focus of CDI is to change the quality of the parent-child 
relationship. Parents are taught to let the child lead 
the interaction, respond to their child enthusiastically, 
listen attentively, describe and praise appropriate 
behavior and ignore (as much as is safely possible) 
inappropriate behavior (McNeil, Eyberg, Eisenstadt, 
Newcomb & Funderburk, 1991; Eyberg, 1979). Special 
emphasis is placed on shaping appropriate child speech. 



27 
Parents are specifically coached to attend to and 
reinforce appropriate child verbalizations by reflecting 

(repeating) the child's statements, answering the child's 
questions, and ignoring negative child verbalizations 

(Eyberg, 1979) . For example, parents are frequently 
coached to ignore "bossy" (direct commands) verbalizations 
made by their children. 

During PDI, parents are taught to continue their use 
of CDI skills as they lead the play. To lead PDI, parents 
are taught how to give simple, direct commands and provide 
consistent consequences for the child's response to each 
command. These consequences include labeled praise for 
compliance and a time-out procedure for noncompliance 

(McNeil et al., 1991; Eyberg & Boggs, 1989). 

Effects of PCIT on Child Talk 

As noted above, PCIT has been shown to decrease 
disruptive behavior, which includes some negative verbal 
behaviors (Eyberg & Robinson, 1982), but to date, no study 
has explicitly examined changes in child verbalizations 
following PCIT. Findings that children fall into the 
normal range of disruptive behaviors following treatment 
and that parent verbalizations become more positive as 
well suggest that children's verbalizations may more 
closely approximate those of normal children following 
PCIT. In the present study, child talk categories were 
coded for 20 behavior problem preschool-age boys with 



28 
their mothers in three DPICS situations prior to and 
following PCIT. Specific hypotheses regarding changes in 
child talk were based on demonstrated differences between 
the child talk of conduct problem children and normal 
children as well as theoretical expectations based on the 
treatment procedures that compromise PCIT (Eyberg & Boggs, 
1989) . 

Specific Hypotheses 
The following hypotheses regarding changes pre- to 
post-PCIT were tested in the present study: 

1) Based on previous differences in deviant 
behaviors between conduct problem and comparison children 
and on the direct coaching of parents in PCIT to ignore 
negative verbalizations including whine, smart talk, yell, 
and critical statements, these child verbalizations will 
decrease following treatment. 

2) Based on findings that conduct problem boys emit 
more commands than comparison children and that parents 
are coached in PCIT to ignore "bossy" verbalizations, 
child commands will decrease following treatment. 

3) Based on findings that conduct problem children 
ask fewer questions in play than comparison children and 
because parents in PCIT are coached to reinforce child 
questions by answering them the first time they occur 
(i.e. immediately), child guestions will increase 
following treatment. 



29 

4) Based on findings that conduct problem children 
engage in more playtalk than comparison children and that 
parents are coached in PCIT to speak directly to their 
children in play and avoid playtalk, conduct problem 
children will model their parent's communication and 
engage in less playtalk following treatment. 

5) Based on the finding that conduct problem children 
emit less praise than comparison children, and that 
parents are being coached to praise their child 
extensively during PCIT, conduct problem children will 
model their parent's communication and themselves emit 
more praise following PCIT. 

6) The last hypothesis was exploratory in nature. An 
implicit goal of PCIT is to increase the mutual give and 
take in the verbal interaction. Parents are coached to 
follow the child's lead and to respond immediately to 
appropriate child verbalizations. For children who tend 
to dominate the verbal interaction, this coaching was 
expected to increase the parent's rate of speech and 
thereby decrease the child's rate. In contrast, for 
children who initially verbalize little, the parent's 
verbalizations were expected to reinforce and thereby 
increase the child's verbalizations. Therefore, it was 
hypothesized that the total quantity of speech of the 
mother and the child would be more similar following 
treatment than it was prior to treatment. 



METHODS 

Subjects 

The subjects were 20 mother-son dyads who had been 
referred to the Shands Hospital Psychology Clinic by 
physicians, mental health professionals, and school 
personnel for the treatment of child behavior problems. 
To be included in this study, children had to have met the 
following criteria at the time of the intake evaluation: 
(1) be between the ages of three years, zero months and 
seven years, zero months; (2) receive parent ratings on 
the Eyberg Child Behavior Inventory (ECBI) above the 
published cut-off scores for child deviancy, that is, an 
Intensity Score > 127 and a Problem Score > 11 (Eyberg & 
Ross, 1978); (3) receive a DSM-III-R diagnosis of either 
Oppositional Defiant Disorder, Attention Deficit 
Hyperactivity Disorder or Conduct Disorder (APA, 1987) . 

Additionally, these dyads had to have completed PCIT 
training with both pre- and post-treatment DPICS 
observations recorded on videotape. Subjects were 
selected by going back in time through the clinic tape 
logs until 20 mother-son dyads were collected who had met 
the above criteria. No parent-child dyads were excluded 
on the basis of race or socioeconomic status. Children 

30 



31 
with moderate to severe mental retardation and children 
with significant language delay or articulation 
difficulties, however, were excluded from the study. 

The sample included 18 white and 2 black dyads. The 
children's mean age was 5.2 years (range 3.2 to 7.0 years) 
at the beginning of treatment. Forty-five percent of the 
children were being raised by single mothers. At 
pre-treatment the mean ECBI Intensity Score was M = 162.55 
(SD = 26.95) and the mean Problem Score was M = 22.95 (SD 
= 6.62). These scores had decreased to within normal 
limits (Eyberg & Ross, 1978) by post-treatment with a mean 
ECBI Intensity Score of M = 105.1 (SD = 21.24) and a mean 
Problem Score of M = 7.3 (SD = 6.1). The length of 
treatment averaged 14.1 sessions (SD = 2.73) with a range 
from 7 to 20 sessions. 

Observational Measure 
Expanded Dyadic Parent-Child Interaction Coding System 

An expanded revision of the Dyadic Parent-Child 
Interaction Coding System (DPICS) (Robinson & Eyberg, 
1981; Eyberg & Robinson, 1983) was used to code all 
interactions. The DPICS is a direct observation procedure 
for assessing the guality of interactions between parents 
and children ranging from 2 to 7 years. It was 
specifically designed to assess treatment progress and be 
used as a pre- and post-treatment observational measure in 
the clinic setting (Eyberg & Robinson, 1983). 



32 
The DPICS was originally designed for coding all 
parent speech using nine categories and several select 
negative child categories (e.g. whine, smart talk, yell) . 
The DPICS coding system was subsequently expanded to code 
all child verbalizations exhaustively, by applying the 
same nine parent verbalization category codes to the 
child's speech. In an initial study (Forster, 1984) the 
child verbalization categories were studied in 20 conduct 
problem and 20 comparison mother-child dyads in the CDI 
situation of DPICS. Results indicated that these 
categories could be reliably coded, and suggested that 
child speech categories could discriminate conduct problem 
from comparison mother-child dyads. 

In DPICS the parent and the child are observed in 
three standard five minute situations: Child-Directed 
Interaction (CDI) ; Parent-Directed Interaction (PDI) ; and 
Clean-Up. Directions are provided to the parent through a 
bug-in-the-ear microphone (Eyberg & Robinson, 1981) . 
In the first situation, labeled Child-Directed 
Interaction (CDI) , the parent is instructed to allow the 
child to choose any activity and to play along with the 
child according to the child's rules. For CDI, the parent 
is instructed as follows: "In this situation, tell 
(child's name) that he/she may play whatever he/she 
chooses. Let him/her choose any activity he/she wishes. 
You just follow his/her lead and play along with him/her". 



33 

In the second situation, labeled Parent-Directed 
Interaction (PDI) , the parent is instructed to choose an 
activity and keep the child playing according to the 
parent's rules. At the end of the CDI situation the 
following instructions are given to the parent to begin 
PDI. "That was fine. Now we'll switch to another 
situation. Tell (child's name) that it is your turn to 
choose the game. You may choose any activity. Keep 
him/her playing with you according to your rules". 

In the third situation, Clean-up, the parent is 
instructed to get the child to put away all the toys 
without assistance. The following instructions were given 
to the parent to begin Clean-up. "That was fine. Now I'd 
like you to tell (child's name) that it is time to put the 
toys away. Get him/her to put all the toys away without 
your help. You should get (child's name) to put each toy 
in its container and then put all the containers into the 
big toy box in the corner. Remember, he/she needs to put 
them away all by him/herself". 

The structure of the situations allows the parent and 
the child to interact naturally with increasing 
requirements for parental control. The three situations 
always are observed in the same sequence, and each 
situation is coded for 5 minutes (Eyberg & Robinson, 1983) . 

The first two situations of the DPICS (CDI and PDI) 
were standardized on 20 families with children referred 



34 
for conduct problem behaviors and 22 normal families 
(Robinson & Eyberg, 1981) . All children in the 
standardization sample were between the ages of two years, 
zero months and seven years, zero months. Norms for the 
parent and child behavior categories are presented in the 
DPICS manual. Mean interrater reliability of .91 (range = 
.67 - 1.0) and .92 (range = .76 - 1.0) were reported for 
the parent and child behavior categories (Robinson & 
Eyberg, 1981) . 

Discriminant function analysis of the DPICS 
observations in the standardization sample revealed 
correct classification of 94% of families according to 
treatment status (Robinson & Eyberg, 1981) . The DPICS was 
also shown to predict 61% of the variance in parent report 
of home behavior problems as measured by the Eyberg Child 
Behavior Inventory (Robinson & Eyberg, 1981) . Sensitivity 
of the coding system to treatment effects has been 
demonstrated by studies showing significant changes in 
parent speech and in child negative talk (Eisenstadt, 
1990; Eyberg & Matarazzo, 1980; Eyberg & Robinson, 1982). 
Specifically, in CDI, mothers show decreases in direct 
commands, indirect commands, questions, and critical 
statements, and increases in descriptive statements and 
total praise; in PDI, mothers show decreases in indirect 
commands and questions, and increases in total praise 
(Eyberg & Matarazzo, 1980; Eyberg & Robinson, 1982). 



35 
Across all three DPICS situations, children show decreases 
in yells, whines, and smart talk (Eisenstadt, 1990) . 

Procedure 
All verbalizations of both the mother and the child 
for the subjects of this study were transcribed by the 
primary investigator from both pre- and post-PCIT 
videotapes. The initial 5 minutes of CDI, PDI and 
Clean-Up were used for all subjects from both pre- and 
post-tapes. Each unit of verbal behavior coded followed a 
"one sentence rule", meaning each verbal unit formed one 
sentence of coding. When verbal phrases ran together in a 
series, a "two second rule" was implemented. That is, 
each time there was a two-second break or pause and then a 
continuation, the verbal phrase after the pause was coded 
as a separate sentence. Lastly, each discrete verbal 
behavior was coded into only one category. A decision 
rule of priority coding determined to which category the 
sentence applied when there was a guestion of a sentence 
fitting into more than one category (see table 1) . No 
punctuation (e.g. guestion marks etc.) was added to the 
sentences on the transcript. 



36 



TABLE ONE 
PRIORITY ORDER FOR CODING CHILD SPEECH CATEGORIES 



Playtalk 

Smart Talk 

Yell 

Whine 

Critical Statement 

Indirect Command 

Direct Command 

Unlabeled Praise 

Labeled Praise 

Descriptive/Reflective Question 

Reflection 

Following Description 

Leading Description 

Acknowl edgment 



Coders in this study were two female graduate 
students trained in DPICS coding and involved in the 
research for the expansion of DPICS to include coding of 
children's speech. Additionally, the coders were 
continuing to use DPICS coding in their own PCIT therapy 
cases on a weekly basis during this study. 



37 
The child verbalizations were then coded by one of 
two trained coders (trained in expanded DPICS) from the 
transcription and simultaneous playback of the videotape. 
Transcriptions were provided to improve the amount of 
child speech that was discernible while the videotape was 
used to provide context and tone inflection. 
Additionally, 3 0% of the parent-child interactions were 
further coded by a third coder (the primary investigator) 
for reliability. This 30% was randomly drawn from all six 
PCIT situations (e.g. CDI pre, PDI post etc.) for adequate 
representation. Reliability of coding was assessed using 
both the Pearson's R correlation between coders by speech 
category as well as sentence-by-sentence inter-coder 
agreement. Additionally, reliability was assessed by use 
of a criterion tape that was intermittently recoded by all 
coders in the study to assess coder drift. 

Child Speech Categories 
The expanded revision of DPICS consists of the same 
three dyadic situations (CDI, PDI, C-U) contained in 
DPICS, in which parent-child interactions are observed. 
The major change involves the inclusion of additional 
child verbalization categories that parallel the parent 
verbalization categories. Mean reliability for the 
expanded category codes of .78 has been reported (Wolf, 
1990) . The 14 child verbalization categories to be coded 



38 
include the following negative and non-negative behaviors: 
Negative 

1. Smart Talk includes impudent or disrespectful 
speech (e.g. "I'm not going to do what you say", any "no" 
following a parental command) . 

2. Yell is a loud screech, scream, or verbalization 
that is clearly above the intensity of normal 
conversation. 

3. Whine is a word uttered by the child in a 
slurring, nasal, high-pitched, falsetto voice. 

4. Critical Statement is a verbalization that finds 
fault with the parent, with the child himself, or some 
object or person unrelated to the ongoing interaction. 
Critical Statements are divided into two categories: 
Critical Statement of Self and Critical Statement of 
Other. 

Nonneqative 

1. Indirect Command is an order, demand, or 
direction for a behavioral response that is implied, 
nonspecific, or stated in a question form. 

2. Direct Command is a clearly stated order, demand, 
or direction in a declarative form; it must be 
sufficiently specific as to indicate the behavior that is 
expected from the parent. 



39 

3. Unlabeled Praise is a verbalization that 
expresses a nonspecific favorable judgment of the parent, 
the child himself, or some object, activity or person 
unrelated to the ongoing interaction. Unlabeled Praise is 
divided into categories: Unlabeled Praise of Self and 
Unlabeled Praise of Other. 

4. Labeled Praise is any verbalization that 
expresses a specific favorable judgment of the parent, the 
child himself, or some object, activity or person 
unrelated to the ongoing interaction. Labeled Praise is 
divided into categories: Labeled Praise of Self and 
Labeled Praise of Other. 

5. Descriptive/Reflective Question is a descriptive 
or reflective comment or acknowledgement expressed in 
guest ion form. 

6. Reflective Statement is a declarative phrase or 
statement which immediately repeats the parent ' s 
verbalization. The reflection may be exactly the same 
words the parent used, may contain synonymous words, or 
may contain some elaboration upon the parent's statement, 
but the basic content must be the same as the parent's 
message. 

7. Descriptive Statement is a declarative sentence 
or phrase that gives an account of the objects or people 
in the situation or the activity occurring during the 
interaction. It can be: (a) a description of the child 



40 
himself, or the child's own activity (b) a description of 
the parent or the parent's activity, or (c) a description 
of an object, person, or activity unrelated to the ongoing 
activity. These descriptive statements are then 
classified as either Leading or Following and coded 
accordingly. A Leading Description is a statement that 
introduces information that is relevant to the play or 
events in the playroom but are not clearly following the 
child's or the parent's behavior or just prior 
verbalization. A Following Description is a statement 
that follows and refers to the child's or parent's 
behavior or just prior verbalization. 

8. Acknowledgment is a brief verbal response that 
indicates attention to the parent's verbal or nonverbal 
behavior but does not describe or evaluate the people in 
the dyad or their behavior. 

9. Playtalk is a verbalization in which the child 
speaks as a character in a game (e.g. speaks like a 
"Cookie Monster") . 



RESULTS 
Reliability 
The primary coders were initially trained on child 
talk categories and following training, reliability was 
assessed against a criterion tape before coding actual 
data in this study. Coder reliability was assessed using 
the Pearson's correlation by child talk category with 
coders averaging .98 reliability (range .97 to .99) 
across coders. Additionally, reliability was assessed 
during coding using the criterion tape which all coders 
intermittently re-coded to assure maintenance of proper 
coding without coder drift. The criterion tape to assess 
coder drift during the actual coding of the data was coded 
on five separate occasions by all coders independently. 
Coders independently re-coded the criterion tape after 
coding approximately 15-25 data tape segments. 
Reliability using Pearson's correlation by child talk 
category across coders averaged .98 (range .96 to .99) for 
all criterion checks. Taped segments were coded by the 
coders in random seguences with the coder kept blind to 
condition (pre vs. post) and situation (CDI, PDI, C-U) to 
minimize consensual coder drift. 



41 



42 
All data for this study were coded from 
transcriptions by one of two primary coders while 
simultaneously viewing the interaction segments on 
videotape. Thirty percent of the segments were also coded 
by a third coder in the same manner to obtain interrater 
reliability. This thirty percent was randomly drawn in 
egual amounts from all six PCIT situations for adequate 
representation. That is, six segments were randomly drawn 
from CDI pre-treatment , six from PDI pre-treatment, six 
from Clean-Up pre-treatment, six from CDI post-treatment, 
six from PDI post-treatment and six from Clean-Up post- 
treatment. Both the Pearson's R correlation and 
percentage agreement were calculated to obtain the 
reliability estimates. 

Percentage agreement for each child talk category and 
for each summary variable was obtained by dividing the 
total frequency of agreements by the total frequency of 
agreements plus disagreements for the two coders. Percent 
agreement of the individual child talk categories is shown 
in table 2. These child talk categories were then 
combined into summary variables to address the hypotheses 
of this study. 

The definitions of these summary variables were as 
follows: Negative Child Talk was a combination of the 
child talk categories Smart Talk, Critical Statements, 
Whine, and Yell; Child Command was a combination of both 






43 
Direct and Indirect Command child talk categories; Child 
Question consisted of the Descriptive/Reflective child 
talk category; Playtalk used the child talk category of 
the same name; and Child Praise was a combination of Child 
Labeled and Unlabeled Praise of Self and Labeled and 
Unlabeled Praise of Other. 

TABLE TWO 
RELIABILITY OF CHILD TALK CATEGORIES 



Category 



Percent Agreement 



Playtalk .56 

Smart Talk .77 

Yell .60 

Whine .78 

Critical Statement .62 

Indirect Command .84 

Direct Command .75 

Praise .75 

Descriptive/Reflective Questions .91 

Reflection .81 

Following Description .64 

Leading Description .84 

Acknowledgment .89 



44 



Description was a combination of Leading and Following 
Description child talk categories and the summary 
variables Acknowledge and Reflection came from the child 
talk categories of the same names respectively. 



TABLE THREE 
RELIABILITY OF SUMMARY VARIABLES 



Variable 



Percent Agreement Pearson's Correlations 



Negative Child Talk 


.87 


Commands 


.89 


Questions 


.91 


Playtalk 


.56 


Praise 


.75 


Description 


.93 


Acknowledge 


.89 


Reflection 


.81 


Total M 


.83 



.97 
.98 
.98 
.83 

1.00 
.99 
.99 

1.00 

.97 



45 
Percent agreement and Pearson's R reliability correlations 
for the summary variables used to test the hypotheses of 
this study are listed in table 3. The mean percent 
agreement for the coding categories ranged from .56 to .93 
and the mean Pearson's ranged from .83 to 1.0. 

Overall interrater agreement by situation for the 
summary variables was .91 (range .68 - 1.0). Interrater 
agreement by situation at both pre and post-treatment are 
shown in table 4. These results are similar in magnitude 
to reliability coefficients for DPICS reported elsewhere 
(Forster, Eyberg & Burns, 1990; Robinson & Eyberg, 1981). 



TABLE FOUR 
ERRATER AGREEM 











PRE 




POST 


SITUATION 







CDI 


.95 


.95 


PDI 


.85 


.88 


Clean-Up 


.93 


.95 



OVERALL M = .91 (Range of individual segments .68 - 1.0) 






46 
Pearson correlations between each primary coder and 
the reliability (third) coder for the five hypothesis 
variables were also performed. Correlations between Coder 
One (coding 66 percent of the reliability segments) and 
the reliability coder were as follows: Negative Child 
Talk = .98, Child Commands = .97, Child Questions = .98, 
Play Talk = .93 and Praise = .84. Correlations between 
Coder Two (coding 33 percent of the reliability segments) 
and the reliability coder were as follows: Negative Child 
Talk = .96, Child Commands = .99, and Child Questions = 
.98. No analyses on the summary variables Praise or 
Playtalk were performed for Coder Two because there were 
no instances of these two categories in this 3 3 percent of 
the reliability tapes. Table 5 shows the Pearson 
Correlations by summary variable for all reliability data 
and for each coder separately. 

Child Talk Variables Hypotheses 
Theoretical considerations and previous research led 
to the investigation of five main hypotheses regarding 
pre- to post-treatment changes in children's speech. Five 
specific child talk summary variables were created by 
combining specific child talk categories. 



47 

TABLE FIVE 
SLATIONS BY SUMMARY VARIABLES 





VARIABLE 


ALL 


CODER ONE 


CODER TWO 




(N=36) 


(N=24) 


(N=12) 


Negative 
Child Talk 


.97 


.98 


.96 


Commands 


.98 


.97 


.99 


Questions 


.98 


.98 


.98 


Playtalk 


.83 


.93 


- 


Praise 


1.00 


.84 


— 



These summary variables were then examined using repeated 
measures analyses of variance to evaluate pre- to post- 
treatment changes among each of the three DPICS 
situations. Five time (pre vs. post) by situation (CDI, 
PDI, C-U) , ANOVA's were performed with these child talk 
summary variables as dependent variables. Where 
significant differences among situations were obtained, 
post-hoc analyses using t-tests were performed. These 
pairwise comparisons, which utilized a Studentized Maximum 
Modulus correction procedure (Dunn, 1974) to control for 



48 
Type I error, were performed to delineate the situations 
(CDI, PDI, C-U) that were significantly different. 

The first variable investigated was Negative Child 
Talk. The hypothesis was that based on the direct 
coaching of parents in PCIT to ignore negative 
verbalizations, these verbalizations would decrease over 
the course of treatment. The analysis revealed no 
significant difference in Negative Child Talk among the 
three PCIT situations at either pre-or post-treatment, F 
(2, 38) = .37. There were also no significant interaction 
effects between time and situation, F (2, 38) = 1.95. 
However, as predicted, Negative Child Talk decreased 
significantly over time from pre- to post-treatment, F (1, 
19) = 13.35, p < .001. Mean rates per situation and 
standard deviations for the Negative Child Talk summary 
variable are presented in table 6. 

The second variable was Child Command. Table 7 shows 
the mean rates per situation and standard deviations by 
situation for this summary variable. Because conduct 
problem boys, when studied in a CDI situation, had been 
shown to emit more commands than comparison children 
(Forster, 1984) , it was hypothesized that the number of 
commands would decrease pre- to post-treatment. However, 
no significant reduction in Child Command was found in 
pre- to post-treatment analyses, F (1, 19) = .72. There 
were also no significant interaction effects between time 



49 



and situation, F (2, 38) = .98. Results of the analyses 
did reveal highly significant differences between 
situations, F = 15.86 (2, 38) p < .0001. 



TABLE SIX 

NEGATIVE CHILD TALK 
MEANS AND STANDARD DEVIATIONS 



PRE 



POST 



SITUATION 


M 


SD 


CDI 


5.47 


(8.96) 


PDI 


7.47 


(6.54) 


Clean-Up 


8.68 


(7.29) 



M 
3.37 
3.10 
1.53 



SD 
(3.06) 
(3.91) 
(1.95) 



Post hoc analyses using pairwise comparisons showed that 
children emitted significantly more commands in CDI than 
in PDI, t (19) = 4.69, p < .01 or Clean-Up, t (19) = 
5.06, p < .01. There were no significant differences 
between PDI and Clean-Up in the number of Child Commands, 
t (19) = 1.59. 



50 

The third variable was Child Question. The 
hypothesis was that teaching parents to answer a child's 
question the first time it was asked would reinforce the 
child for asking questions and thereby lead to an increase 
in the frequency of child questions from pre- to post- 
treatment . 



TABLE SEVEN 

CHILD COMMANDS 
MEANS AND STANDARD DEVIATIONS 



PRE 



POST 



SITUATION 


M 


SD 


CDI 


7.26 


(5.86) 


PDI 


3.05 


(4.99) 


Clean-Up 


2.16 


(2.89) 



M 


SD 


5.00 


(4.54) 


3.16 


(3.53) 


2.05 


(2.12) 



There were no significant differences in Child 
Question among the three situations, F (2, 38) = 2.08. 
There were also no significant interaction effects between 
time and situation, F (2, 38) = .35. Contrary to the 



51 
proposed hypothesis, there was a significant reduction in 
the frequency of the Child Question variable following 
treatment, F (1, 19) = 7.12 p < .01. Table 8 shows mean 
rates per situation and standard deviations by situation. 
The fourth variable was Playtalk utilizing the child 
talk category of the same name. The total amount of 
Playtalk in this sample was quite low, and because of the 
low reliability between coders on this category, results 
of these analyses can only be viewed with caution. 



TABLE EIGHT 

CHILD QUESTIONS 
MEANS AND STANDARD DEVIATIONS 



PRE POST 



SITUATION M SD M SD 

CDI 6.58 (5.19) 3.63 (3.89) 

PDI 7.16 (7.02) 5.16 (3.65) 

Clean-Up 6.58 (5.60) 3.53 (2.87) 



52 
In an earlier study (Forster, 1984) , playtalk was 
found to occur significantly more often during CDI among 
behavior problem than comparison children. The hypothesis 
was that PCIT would result in decreased playtalk across 
all situations post-treatment. Analyses revealed no 
significant differences between pre- and post-treatment 
scores on this variable, F (1, 19) = .47, and no 
significant interaction effects, F (2, 38) = .12. The 
difference between situations was significant, F (2, 38) = 
4.86, p < .01, but post hoc analyses revealed that 
Playtalk in the CDI situation as compared to the PDI 
situation only approached significance, t (19) = 2.41 , p 
= .03, and was not significantly different at the .01 
level. Post hoc comparisons revealed no significant 
differences between PDI and Clean-Up, t (19) = 1.16 and 
between CDI and Clean-Up, t (19) = 2.04. Table 9 lists 
mean rates per situation and standard deviations by 
situation for Playtalk. 

The fifth and final child talk variable investigated 
was Child Praise. The hypothesis was that as parents were 
being taught in PCIT to increase their use of praise that 
some modeling would occur resulting in an increase in the 
child's use of praise. Results of the analysis of 
variance showed no significant differences across time, F 
(1, 19) = .16, across situation, F (2, 38) = 1.54 or in 
interaction, F (2, 38) = .85. Again, as in Playtalk, the 



53 



frequency of Child Praise was very low. Mean rates per 
situation and standard deviations are presented in table 
10. 



TABLE NINE 

CHILD PLAYTALK 
MEANS AND STANDARD DEVIATIONS 



PRE 



POST 



SITUATION 



M SD 



M 



SD 



CDI 
PDI 
Clean-Up 



1.26 (4.57) 
.10 ( .31) 
.21 ( .71) 



1.84 (2.99) 
.16 ( .37) 
.42 (1.17) 



Total Speech Hypothesis 
The final hypothesis of this study examined the total 
quantity of speech of both the mother and the child at 
pre- and post-treatment. It was hypothesized that the 
amount of speech for mother-child dyads would become more 
similar from pre- to post-treatment as a result of the 









54 
dyad developing better reciprocity in communicating. 
Amount of speech was defined as the number of sentences 
for both the mother and the child as transcribed from the 
videotapes. Sentences were defined according to the DPICS 
rules described in the method section. The data were 
analyzed using a Pearson's correlation analysis comparing 
the difference between the mother's and the child's amount 
of speech at pre-treatment to the difference between the 
mother's and the child's amount of speech at post- 
treatment . 



TABLE TEN 

CHILD PRAISE 
MEANS AND STANDARD DEVIATIONS 



PRE 



POST 



SITUATION 

CDI 

PDI 

Clean-Up 



M 


SD 


42 


(.90) 


10 


(.31) 


05 


(.23) 



M 


SD 


.26 


(.56) 


.21 


(.71) 


.21 


(-42) 



55 
Contrary to the hypothesis, the amount of difference 
between mother-child dyads at pre-treatment was not 
related to the amount of difference in the dyads at post- 
treatment. Thus, some dyads became more similar and some 
became more different with the overall difference between 
mother-child dyads becoming larger at post-treatment, 
though not significantly so. Additionally, when the 
difference between the dyadic mean score at pre-treatment 
was compared to the dyadic mean score at post-treatment 
using a T-test, (pre M = 177.9; post M = 152.0) this 
difference was also not significant, t(19) = 1.37. From 
pre- to post-treatment there was a significant diffrence 
with children speaking less at post-treatment, t (19) = 
4.67, p < .001. Table 11 shows means, standard deviations 
and pre to post correlations for this hypothesis. 



TABLE ELEVEN 
TOTAL SPEECH MOTHER TO CHILD 



MOTHER CHILD CORRELATION DIFFERENCE 

SCORE 



M SD M SD M SD 

PRE 219.4 (51.1) 136.4 (37.0) .39 82.9 (50.0) 

POST 206.5 (51.7) 97.6 (39.2) .32 108.9 (53.6) 



DISCUSSION 

The purpose of this study was to compare the 
verbalizations of behavior problem boys before and after a 
treatment program designed to address psychological 
disturbances in pre-school age children. The therapy 
program, Parent-Child Interaction Therapy, changes the 
overt behaviors of both the child and the child's primary 
caretakers while additionally teaching the primary 
caretakers to decrease their negative verbalizations and 
increase their positive verbalizations to their child. 
What had not been investigated until this study was the 
subsequent effect the therapy program would have on the 
child's verbalizations. 

The basic hypotheses were (a) that over the course of 
the therapy program the child's negative verbalizations 
would decrease and pro-social verbalizations would 
increase; and (b) that these changes would lead to better 
give-and-take between the mother-child dyad as reflected 
by a greater similarity in the number of verbalizations 
within each dyad following treatment. The results suggest 
that the therapy program does indeed affect the child's 
verbalizations in a positive way. Results indicate 

56 



57 
significantly fewer negative verbalizations after 
treatment from these children. However, some speech 
categories changed in ways not predicted and the total 
amount of speech within the mother-child dyads actually 
became more dissimilar rather than more similar. 

Reliability 
Several interesting findings came from the 
reliability coding which will be discussed first. 
Analyses of the percent agreement for the individual child 
talk categories, indicated that several received marginal 
reliability ratings. Specifically, Following Description, 
Critical Statement, Yell and Playtalk were all below .75 
in reliability. In the case of the category labeled 
Following Description, it appears that coders had 
difficulty discriminating following from leading rather 
than a description versus another category because when 
combined with Leading Description and re-analyzed the 
percent agreement for Description increased to .93. 
Likewise, for Critical Statement and Yell, when these two 
categories were included in the summary variable Negative 
Child Talk, the percent agreement increased to .87. It 
appears that having to make judgements about volume and 
tone were significant contributors to difficulty in coding 
these categories reliably although coders did agree that 
the category should be a Negative Child Talk of some kind. 



58 
Tone, inflection and volume continue to demand a great 
deal of attention in refining coding systems so as to 
operationalize these qualitative aspects of negative 
speech more definitively and provide reliable distinctions 
among the discrete categories. 

For Playtalk, it appears now that the category itself 
was not defined specifically enough before coding began 
and due to its sparse appearance this was not discovered 
during training of coders. It seems that coders were 
unclear whether certain unexpected verbal behaviors, such 
as singing, should be coded as playtalk. 

For the summary variables, the Pearson's correlations 
were found to be higher for all categories than the 
percent agreement results. The Pearson's correlation 
provides information on the extent to which two coders' 
scores show similar patterns across a session or block of 
time. Yet this method is often less sensitive to bias 
across coders, does not yield information about coder 
agreement on specific occurrences, and has been reported 
to lead to possible overestimation of agreement (Johnson & 
Bolstad, 1973) . Overestimation can happen when one coder 
consistently rates all behaviors higher than a second 
coder, and thus Pearson's correlation may inflate 
reliability estimate between coders to some extent. 
Therefore, the more stringent interval or percent 
agreement was used as well. 



59 
The percent agreement is often considered more 
stringent than the Pearson's correlation because of its 
focus on discrete intervals (Poling & Fuqua, 1986) . In 
order to obtain high agreement scores, coders must agree 
on the occurrence of the target behavior at each 
particular interval. This tends to increase the 
probability that they are coding the same behavior at the 
same moment. However, percent agreement can be negatively 
affected by extreme rates in the dependent variable such 
as the very low occurrences in this study of Playtalk and 
Child Praise, and percent agreement has been criticized 
for its lack of metric properties which limits its further 
use in analyses. 

As a result, both the Pearson's correlation and the 
percent agreement were considered important for obtaining 
estimates of reliability. As complimentary techniques 
they were able to assess reliability on a situation basis 
as well as on a discrete sentence-by-sentence basis. 
Among the percent agreement scores, lower reliabilities 
indicate more definitively the coding categories that were 
more difficult to code accurately and which may need 
increased operationalization. 

One more issue appeared from the reliability 
analyses. When interrater agreement was analyzed by 
situation, a slight drop in agreement was found in the PDI 
situation in comparison to the CDI and Clean-Up 



60 
situations. It has been shown in previous research 
(Eyberg & Robinson, 1982) that children emit increased 
incidence of deviant behaviors (including negative verbal 
behaviors) in the PDI as opposed to the CDI situation of 
DPICS. It may have been this increase in deviant verbal 
behavior, which includes more codes which are 
significantly influenced by tone and volume, that made 
reliable coding of PDI more difficult to attain. However, 
the reliability in Clean-Up was higher than PDI despite 
having a similar amount of negative child talk when pre 
and post conditions were combined. What may be more 
likely is that the lower reliability of PDI is the result 
of the tape segments being more difficult to code in PDI 
due to the actual behavior of the subjects randomly drawn 
or due to something implicit in the PDI situation that 
made coding more difficult. 

Child Talk Summary Variable Hypotheses 
The first main child talk summary variable 
investigated was Negative Child Talk. Consistent with the 
expectation, PCIT significantly lowered the amount of 
negative verbalizations in behavior problem boys due to 
the direct coaching of mothers in PCIT to ignore negative 
verbalizations. This is an especially important finding 
given that these children's negative verbalizations, 
including those verbalizations emitted with a negative 



61 
tone or volume component, are often major contributors to 
their interpersonal difficulties. Analyses revealed no 
significant difference at pre-treatment between the three 
situations although other studies (Forster, 1984) have 
obtained higher amounts of negative child talk in Clean-Up 
than in CDI or PDI. 

Another area of difficulty for some behavior problem 
children appears to be the degree to which they attempt to 
direct the mother-child interaction with Commands, which 
are defined as verbalizations that tell or request the 
mother to perform some behavior. Previous research 
(Forster, 1984) has shown that behavior problem boys give 
commands to their mothers during CDI at a higher rate than 
comparison boys. In analyzing the Child Command summary 
variable across the three situations, however, no decrease 
over the course of treatment was found. It is possible 
that the pre-treatment commands of behavior problem boys 
become more "bossy" in the PDI and Clean-Up situations, 
along with the increase in negative talk categories, 
leading to a change in tone or pitch such that these 
verbalizations are coded Smart Talk rather than Commands. 
This qualitative change could result in fewer "commands" 
being coded in PDI and Clean-Up, thus decreasing their 
overall rate across the three situations. In fact, the 
post hoc analyses of Child Command indicates that they are 



62 
highly situation specific with many more occurring in CDI 
than in PDI or Clean-Up. 

In CDI the parent is instructed to tell the child 
that s/he can play any game they want and the parent's 
task is to follow along and play by the child's rules. 
The finding of more Commands in CDI most probably is a 
direct result of the instructional set that limits the 
parents' control of the situation and maximizes the 
child's sense of control. In a sense, the child is 
reinforced for giving commands by the mother's quick 
compliance. 

The third summary variable investigated was Child 
Question. In contrast to expectations, the amount of 
Child Questions actually decreased from pre- to post- 
treatment. During PCIT, parents are taught to ignore all 
deviant child verbalizations while responding to all 
prosocial child verbalizations including questions. It 
was hypothesized that this differential attention would 
result in the reinforcement of child questions leading to 
a subsequent increase over the course of treatment. 
However, the obtained findings are perhaps explainable in 
terms of positive treatment outcome in that as parents are 
taught to attend more closely to their child and answer 
the child's question the first time, they may be 
eliminating the child's need to repeat a question several 
times to receive a response. Future research would need 



63 
to examine the content of the child's questions to be able 
to delineate whether or not they are merely repetitive or 
new and different questions that the child is asking. 

The final two specific speech category hypotheses 
involved Child Praise and Playtalk. For both of these 
variables the actual number of occurrences was quite low, 
and this study found no significant changes for either 
category from pre- to post-treatment. Although a previous 
study (Forster, 1984) had demonstrated that during CDI 
behavior problem children used praise significantly less 
than a comparison group and used playtalk significantly 
more than the same comparison group, this study did not 
find changes that were predicted based on these results. 
That is, at post-treatment it was expected that the 
behavior problem boys in this study would show a decrease 
in the amount of playtalk and an increase in praise. 

The PCIT research (Eyberg & Matarazzo, 1980) has 
demonstrated that parents increase the amount of praise 
they use over the course of treatment and Eisenstadt 
(1990) found a specific increase in child self-praise 
following PCIT which was one of the speech categories 
incorporated into Child Praise for this study. That no 
change in the amount of Child Praise was found at post- 
treatment in this study may be a consequence of the 
overall low rate of this behavior. 



64 
Although the category Playtalk was found to occur 
more frequently in CDI than PDI, this difference was 
marginally significant. Further, these results must be 
viewed cautiously due to the low coder reliability 
achieved for Playtalk in this study and its low rate of 
occurrence. In the Forster (1984) study, the mean 
frequency of Playtalk in the CDI situation was 3.6 
utterances for conduct problem boys whereas in this study 
it was 1.26 per situation. Overall, the results from this 
study suggest that this child talk category requires 
further definition, precision and study before its 
usefulness in the evaluation of conduct problem children 
can be determined. 

Comparison of Total Amount of Speech of Mother-Child Dyads 

The hypothesis that the dyads would become more 
similar in the quantities of their speech was not 
confirmed. In this study both the children and the 
mothers used fewer statements at post-treatment than they 
had at pre-treatment (M of dyad at pre-treatment = 177.9, 
M of dyad at post-treatment = 152.0) with the children's 
decrease highly significant. Perhaps the decrease in 
statements for both the mother and child reflects a 
positive treatment outcome as other studies have shown 
that hyperactive children and their mothers tend to speak 



65 
more frequently than normals (Barkley, Cunningham & 
Karlsson, 1983) . 

In the Barkley et al.,(1983) study, the speech of 18 
hyperactive boys and their mothers was compared to a 
control group of 18 boys and their mothers (mean age for 
the boys was 8.8 years). Barkley et al. (1983) reported 
that the hyperactive children and their mothers spoke more 
frequently than the normal children and their mothers in a 
free play situation. In the second phase of the Barkley 
et al. (1983) study the hyperactive children were given 
either a placebo or a dose of methylphenidate and observed 
again. When their frequency of utterances was again 
analyzed (coders were blind to placebo vs. medication 
condition) , the hyperactive boys on medication and their 
mothers showed a significant decrease in their frequency 
of speech. The authors hypothesized that it was the 
hyperactive child's high frequency of utterances that had 
initially increased the mother's frequency, and that as 
the medication caused the decrease in the child's 
frequency of utterances, the mother's rate decreased as 
well. That is, they hypothesized that the child's high 
frequency of utterances had actually caused the mother's 
similarly high rate of utterances. 

Although the children in this study were not part of 
a medication study, they had participated in a treatment 
program (PCIT) that had decreased their disruptive 



66 
behavior, as evidenced by their ECBI scores which fell 
within normal limits at post-treatment. Perhaps this 
caused a similar decrease in the child's frequency of 
utterances which led to a drop in the mother's rate as 
well. This would be consistent with the training the 
parents receive in PCIT to attend to their child and 
respond to all appropriate verbalizations. As discussed 
with the child talk category Questions, this attention may 
lead to less talk on the part of both the child and the 
parent which may be a better indicator of the dyad being 
in synchrony than the original hypothesis. 

Implications for Treatment 
It is believed that this study offers important 
findings to the clinician and/or researcher utilizing 
PCIT. As has been stated, children learn language skills, 
self-regulation and social competence through their 
interactions with caretakers and more capable peers 
(Vygotsky, 1962; Damon, 1983). And it is likewise through 
this same interaction process that children undoubtedly 
learn less desirable behaviors (e.g., noncompliance, 
whining, critical statements) . 

Clinicians in recent years have begun to recognize 
their limitations in terms of time that they can spend 
with a particular child in treatment as opposed to the 
time a caretaker can potentially spend with the same 



67 
child. This has led to an increase in the number of 
parent-training programs in an attempt to solicit the 
caretaker's assistance to effect more lasting change in 
these children. Additionally, it behooves the clinician 
to consider these dyadic systems for treatment as the 
dyad's interactions may have been a major contributor to 
the difficulties which initiated intervention. 

The PCIT offers one such systematic program for 
changing the behavior of both members of the dyad. 
Parent-Child Interaction Therapy research (Eisenstadt, 
1990) has shown that in fact, not only do mothers learn 
different and more positive ways of handling their child's 
noncompliance but their verbal skills change as well over 
the course of treatment. Mothers have been shown to use 
fewer Critical Statements, Indirect Commands and Questions 
following PCIT treatment and to increase the amount of 
Praise they use (Eyberg & Matarazzo, 1980) . This study 
has now shown that children's verbalizations change as 
well over the course of treatment. Perhaps this is the 
result of directly teaching the mother to respond to the 
child in a different way (differential reinforcement) 
which replaces the deficient verbal behavior model 
(Patterson et al., 1975) with a more appropriate model. 
This decrease in negative verbalizations points to a 
learning of more appropriate social skills which in turn 



68 
may significantly impact these children's long-term 
prognosis. 

Future Directions 

In general, these findings are encouraging in their 
demonstration of positive changes in child verbalizations 
following a parent-child training program (PCIT) . It is 
also encouraging to note that this measure of child talk 
(expanded DPICS) appears able to capture some of the 
positive as well as the negative categories of change in 
children's speech. 

These findings also indicate the need for further 
research in this area. It will be important to continue 
work on refining the child talk categories of the coding 
system until high reliability can be achieved for those 
categories which are significantly influenced by volume 
and tone. It will also be important to replicate these 
findings with another sample of children and to include an 
untreated control group, to be confident that these 
changes represent more than statistical regression. 

Finally, it will be important to investigate 
children's verbalizations with their fathers to look for 
differences and similarities with each parent, and to 
evaluate the generalization of all these findings to the 
home and other settings. 



REFERENCES 



American Psychiatric Association (1987) . Diagnostic and 
statistical manual of mental disorders ; (3rd 
ed., rev.). (DSM-III-R) . Washington, DC: Author. 

Aragona, J. A., & Eyberg, S. M. (1981). Neglected children: 
Mother's report of child behavior problems and 
observed verbal behavior. Child Development . 52, 
596-602. 

Barkley, R. A., Cunningham, C. E., & Karlsson, J. (1983). 
The speech of hyperactive children and their mothers: 
Comparison with normal children and stimulant drug 
effects. Journal of Learning Disabilities . 16, (2), 
Feb. 

Baum, C.G., & Forehand, R. (1981). Long-term follow-up of 
parent training by use of multiple outcome measures. 
Behavior Therapy , 12, 643-652. 

Bellack, A. S., & Hersen, M. (Eds.). (1979). Research and 
practic e in social skills training . New York: Plenum 
Press. 

Berk, L. E. (1985, July). Why children talk to 
themselves. Young Children , pp. 46-52. 

Brainerd, C. J. (1978) . Learning research and Piagetian 
theory. In L. S. Siegel & C. J. Brainerd (Eds.), 
Alternatives to Piaget : Critical essays on the 
theory (pp. 69-110) . New York: Academic Press. 

Bruner, J. (1975) . The ontogenesis of speech acts. 
Journal of child language . 2, 1. 

Chomsky, N. (1965) . Aspects of the theory of sy ntax. 
Cambridge: MIT Press. 

Damon, W. (1983) . Social and personality development : 

Infancy through adolescence . New York: W. W. Norton 
Press. 



69 






70 

Dunn, 0. (1974) . On multiple tests of confidence 

intervals. Journal of Communications in Statisitics . 
3, 101-103. 

Durkin, K. (1986), (Ed.) Language development in the 
school years. Croom Helm Ltd . , London . 

Eisenstadt, T. (1990) . Parent-child interaction therapy 
with behavior disordered children . Unpublished 
doctoral dissertation, University of Florida. 

Eyberg, S. M. (1979, April). A parent-child interaction 

model for the treatment of psychological disorders in 
young children . Paper presented at the meeting of 
the Western Psychological Association, San Diego. 

Eyberg, S. M. , & Boggs, S. R. (1989). Parent training for 
oppositional-defiant preschoolers. In C. E. Schaefer 
& J. M. Briesmeister (Eds.). Handbook of parent 
training: Parents as cotherapists for children's 
behavior problems . (pp. 105-132). New York: Wiley. 

Eyberg, S. M. , & Matarazzo, R. G. (1980). Training 
parents as therapists: A comparison between 
individual parent child interaction training and 
parent group didactic training. Journal of Clinical 
Psychology . 36 (2), 492-499. 

Eyberg, S. M. , & Robinson, E. A. (1982). Parent-child 

interaction training: Effects on family functioning. 
Journal of Clinical Child Psychology , 11, 130-137. 

Eyberg, S. M. & Robinson, E. A. (1983). Dyadic 

parent-child interaction coding system: A manual. 
Psychological Documents . 13, 24. (Ms No. 2582) 

Eyberg, S. M. , & Ross, A. W. (1978). Assessment of child 
behavior problems:: The validation of a new 
inventory. Journal of Clinical Child Psycholog y. 7, 
113-116. " 

Forster, A. A. (1984). Self-esteem and conduct problems: 
Sex differences and treatment implications 
(Doctoral dissertation, University of Oregon, 1984) . 
Dissert ation Abstracts International r 45 . 1012B. 

Hanf , C. (1969) . A two-stage program for modifying 
maternal controlling during mother-child (M-C^ 
interaction. Paper presented at the meeting of the 
Western Psychological Association, Vancouver, B.C. 



71 

Johnson, S. M. & Bolstad, 0. D. (1973). Methodological 
issues in naturalistic observations: Some problems 
and solutions for field research. In L. A. 
Hammer lynck, L. C. Handy, & E. J. Mash (eds.). 
Behavior Change; Methodology, Concepts, and Practice . 
Champaign, IL: Research Press. 

Kazdin, A. E. (1979) . Sociopsychological factors in 
psychopathology. In A. S. Bellack & M. Hersen 
(Eds.). Research and practice in social skills 
training , (pp. 41-74) . New York: Plenum Press. 

Kazdin, A. E. (1985) . Treatment of antisocial behavior in 
children and adolescents . Homewood; IL: Dorsey 
Press. 

Lenneberg, E. H. (1967) . Biological foundations of 
language . New York: John Wiley & Sons, Inc. 

Lewis, M. , & Cherry, L. (1977). Social behavior and 

language acquisition. In M. Lewis & L. A. Rosenblum 
(Eds.). Interaction, conversation, and the 
development of language , pp. 413-522 New York: John 
Wiley and Sons. 

Lytton, H. (1980) . Parent-child interaction: The 

socialization process observed in twin and singleton 
families . New Yrk: Plenum Press. 

McNeil, C. B., Eyberg, S., Eisenstadt, T. H. , Newcomb, K. , 
& Funderburk, B. (1991) . Parent-child interaction 
therapy with behavior problem children: 
Generalization of treatment effects to the school 
setting. Journal of Clinical Child Psychology . 20 , 
(2), p. 140-151. 

Mead, M. (1963) . Socialization and enculturation . 
Current Anthropology, 4, 184-188. 

Oleron, P., & Lorion, R. (1977). Language and mental 

development (R. Lorion, trans.). Hillsdale, NJ: L. 
Erlbaum Assoc. 

Pascuel-Leone, J. (1970) . A mathematical model for the 
transition rule in piaget's developmental stages. 
Acta psychologica , 32, 301-345. 

Patterson, G. R. (1982) . Coercive family process . 
Eugene, OR: Castalia. 

Patterson, G. R. (1986) . Performance models for 

antisocial boys. American Psychologist . 41, 432-444. 






72 

Patterson, G. R. , Reid, J. B. , Jones, R. R. , & Conger, R. 
E. (1975) . A social learning approach to family 
intervention: families with aggressive children . 
Eugene, OR: Castalia. 

Peed, S., Roberts, M. , & Forehand, R. (1977). Evaluation 
of the effectiveness of a standardized parent 
training program in altering the interaction of 
mothers and their noncompliant children. Behavior 
Modification . 1, 323-350. 

Pelligrini, A.D. (1980). A semantic analysis of 

pre-schooler's self-regulating speech. International 
Journal of Psycholinguistics , 7-4 (20), 59-74. 

Peterson, G. A., & Sherrod, K. B. (1982). Relationship of 
maternal language to language development and 
language delay of children. American Journal of 
Mental Deficiency . 86 (4), 391-398. 

Poling, A. & Fuqua, R. W. (Eds.). (1986). Research methods 
in applied behavior analysis: Issues and advances . 
Plenum Press: New York. 

Robinson, E.A., & Eyberg, S. M. (1981). The dyadic 
parent-child interaction coding system: 
Standardization and validation. Journal of 
Consulting and Clinical Psychology , 49, 245-250. 

Rogoff, B., & Wertsch, J. V. (Eds.). (1983). Children's 
learning in the "zone of proximal development. " San 
Francisco: Josey Bass. 

Rondal, J. A. (1985). Adult-child interaction and the 

process of language acguisition . New York: Praeger 
Publishing. 

Skinner, B.F. (1957). Verbal behavior . New York: 
Appleton-Century-Crof ts . 

Snow, C. E. (1972). Mother's speech to children learning 
language. Child Development . 43 . 549-565. 

Snow, C. E. (1979) . The role of social interactions in 
language acquisition. In W. A. Collins, (Ed.). 
Children's language and communication : The Minnesota 
Symposia on Child Psychology . 12 . Erlbaum Associates, 
Publishers, Hillsdale, N.J. 



73 



Snow, C. E., & Ferguson, C. A. (1977). Talking to 

children: Language input and acquisition . Paper 
presented at a conference sponsored by the committee 
on sociolinguistics of the social science research 
council (USA) . New York: Cambridge University 
Press. 

Thomas, R. M. (1985) . Comparing Theories of Child 

Development . Belmont, CA: Wadsworth Publishing Co. 

Vygotsky, L. S. (1962). Thought and language . Cambridge, 
MA: MIT Press. (Original published 1934) 

Vygotsky, L. S. (1978). Mind in society; The development 
of higher psychological processes . Cambridge: 
Harvard University Press. 

Vygotsky L. S., & Luria, A. R. (1930). The function and 

fate of egocentric speech . Proceedings of the Ninth 
International Congress of Psychology, pp. 17-21. 
Princeton, NJ: Psychological Review. 

Wertsch, J.V. (1979) . The regulation of human action and 
the given-new structure of private speech. In G. 
Zivin (Ed.), The development of self-regulation 
through private speech , (pp 1-22) . New York: Wiley. 

Wolf, R. (1990) . Interobserver reliability as a function 
of coding system complexity . Unpublished Senior 
Thesis. University of Florida, Gainesville, FL. 



BIOGRAPHICAL SKETCH 

Laura Lynn Mee was born in Trumbull, Connecticut, and is 
the oldest child of Frank and Janice Mee. She moved with her 
family from Connecticut to Florida at age 11 and graduated 
from Sarasota High School in 1981. She completed her 
bachelor's degree in psychology in 1985 from the University of 
Florida and obtained a master's degree in clinical psychology 
from Illinois State University in 1987. After completing her 
internship year at Emory University, she will graduate from 
the University of Florida in August, 1991, with a Ph.D. in 
clinical psychology. In July she will begin a postdoctoral 
fellowship at Egleston Children's Hospital at Emory University 
in Atlanta, Georgia, where she resides with her husband and 
cat. 



74 






I certify that I have read this study and that in my 
opinion it conforms to acceptable standards of scholarly 
presentation and is fully adequate, in scope and quality, as 
a dissertation for the degree of Doctor of Philosophy. 



Sheila M. Eyberg, <phair 
Professor of Clinical and 
Health Psychology 



I certify that I have read this study and that in my 
opinion it confirms to acceptable standards of scholarly 
presentation and is fully adequate, in scope and quality, as 
a dissertation for the degree of Doctor of Philosophy. 



Stephen R. Boggs 77 
Assistant Professor of 
Clinical and Health 
Psychology 



I certify that I have read this study and that in my 
opinion it conforms to acceptable standards of scholarly 
presentation and is fully adequate, in scope and quality, as 
a dissertation for the degree of Doctor of Philosophy. 




2^ 



<^L 



Algina) 
isor of Fo\indatioi 
of Education 






I certify that I have read this study and that in my 
opinion it conforms to acceptable standards of scholarly 
presentation and is fully adequate, in scope and quality, as 
a dissertation for the degree of Doctor of Philosophy. 



*^A.K 




Scott A. Miller 
Professor of Psychology 



_ I certify that I have read this study and that in my 
opinion it conforms to acceptable standards of scholarly 
presentation and is fully adequate, in scope and quality, as 
a dissertation for the degree of Doctor of Philosophy. 



Michael J"i Farrar 
Assistant Professor of 
Psychology 



This dissertation was submitted to the Graduate Faculty 
of the College of Health Related Professions and to the 
Graduate School and was accepted as partial fulfillment of 
thee requirements for the degree of Doctor of Philosophy. 

August, 1991 fi r\ u 

lX^ki.\^X ^ 

Dean, College of Health 
Related Professions 



Dean, Graduate School 



UNIVERSITY OF FLORIDA 



3 1262 08554 7221